"How Predictive Analytics ERP Supports You in the Company"

Our guest author Florian Langer has delved into the topic of Predictive Analytics ERP and brings you up to speed on this subject in this article.

Table of Contents

1. What is understood by Predictive Analytics?
2. Who needs Predictive Analytics ERP and why?
3. What does all this have to do with my business now?
   1. How do we plan our production?
   2. How can one look into the future?
4. What ERP systems can I use Predictive Analytics with?
5. Predictive Analytics ERP made in Germany
6. What other ERP systems offer Predictive Analytics?

What is understood by Predictive Analytics?

As a term, artificial intelligence is not entirely uniform to define, partly because it has been developing as an interdisciplinary research direction since the 1950s and has always adapted to the technical possibilities. In general, one can speak of artificial intelligence when an IT system shows "human-like", intelligent behaviors.

Certain core skills are required for this in different proportions, Perception, Understanding, Acting, and Learning. While traditional IT systems focus onInput>Processing> Output, artificial intelligence extends these capabilities to the aspects ofLearningandDoing.

In addition to pure data processing, your system can also learn and be trained during the processing phase. Ideally, a trained setup then achieves better results than the traditional, static methods, which are based on rigid, clearly defined, and firmly programmed rules.

Perhaps you know the Facebook algorithm that optimizes your audience or the placement of your ads towards conversions, that you specify? It's the same principle.

When we look at the individual parts in a little more detail, it becomes clear what potential lies in the use of artificial intelligence. This is due, among other things, to the variety ofData points, that we have available orperceivecan.

Even in small companies applies: “We collect data, create processes and ultimately have to manage them. Classic Excel spreadsheets usually can no longer adequately reflect this amount of data and increase the risk of errors.”

During theData processing come methods like deep learning, language processing or image recognition. Your system is thus extended by a trainable, learning component. For example, you take a photo of your phone bill and your tool can immediately assign it correctly, but if an unstructured request comes in via your chat, your tool/model first has to understand what the customer wants. Alexa from Amazon, for example, does a good job of understanding and distinguishing whether you want to turn on the light in the kitchen or the bathroom. Replika.ai has created a digital buddy that remembers things about you and asks you how you are.

The output componennt of Your Predictive-Analytics-ERP-Systems is the result of perception and processing. During my Data Science boot camp, for example, I was able to pitch and implement a project that translates landscape photos into the painting style of Césanne, Monet, van Gogh or Ukiyo-e. Unfortunately, in practice (and depending on the job), things like an intelligent process or production control are more relevant application areas. In the above example with the invoice, it is then automatically filed correctly and paid properly or further processed.

What is special about these algorithms is that they can learn from feedback during the Training phase but also during ongoing operation and are able to optimize their actions and their output accordingly. The phone bill always comes on the same day, your ERP system remembers that and gives you a hint if there is likely not to be enough money on your AMEX on the due date (this is where the “predictive” from Predictive Analytics ERP comes into play).

Who needs Predictive Analytics ERP and why?

The acronym ERP stands for Enterprise Resource Planning and is a system that can support the automation and management of business processes in the areas of finance, production, sales, marketing, supply chain, human resources and operation. ERP systems help you generate new insights, optimize processes, plan scenarios and improve decision making by breaking down data silos and integrating information between different departments. An ERP system is sometimes referred to as the “central nervous system of a business” as it delivers the automation, integration and intelligence essential for the efficient operation of all daily business processes. Ideally, most or all of your company's data is centralized in the ERP system to have a central data source for the whole company. Especially in the area of conversions and tracking, having a “single source of truth” is always an advantage.

ERP systems can therefore give you a comprehensive overview of the processes and operations in your business and connect the various processes and teams, for example, marketing and product, when it comes to funnel performance. Consequently, you have the opportunity to zoom in on certain sub-processes or to look at your entire company. The overview helps you identify problems, clear obstacles and roadblocks, and make the changes necessary for the further performance of your venture.

A finance team can make accounting more efficient or ensure that suppliers are paid correctly and on time by using a Predictive Analytics ERP tool. In the sales area, you can manage your entire pipeline and score leads. For your marketing team, it is probably also interesting to know how those who consume your content then behave in terms of revenue at the end. Your logistics depends on a well-functioning ERP software to deliver the right products and services to customers on time. On a management level, a real-time overview is certainly beneficial when it comes to making decisions in a timely manner. And finally, you may have better cards with your bank or your investors if you have accurate financial records and forecasts that can be provided by the ERP system.

The importance of ERP software for companies is illustrated by the growing distribution. According to G2 “the global ERP software market is expected to grow to 78.40 billion US dollars by the year 2026, with an average growth rate of 10.2% between the years 2019 and 2026.”

What does all this have to do with my business now?

Let's take a little detour. We have watched this Video from Bill Ackman, but don't have a lemonade stand but a Toy factory. Nevertheless, we face the challenge of manufacturing products with strong seasonal demand.

How do we plan our production?

One option is to produce a fraction of the expected demand every month before the peak season. However, if demand is uncertain, this approach can lead to large surpluses or shortages, which reduce profit margins. Companies in this position must weigh up the costs of increasing production to meet demand during peak season, the additional costs of warehousing and the risk of uncertain demand. But how?

Using the example of Play Time Toy Co., I would like to give an insight into what the engine room of a Predictive-Analytics-ERP system might look like. In particular, I use a Monte-Carlo simulation for this example in order to evaluate the impacts of the production planning on profitability, the capital requirement and, in general, the unit economics of the company.

Play Time Toy Co. is a toy manufacturer facing increasing competition from foreign companies, influencers starting with dropshipping, and Etsy DIY shops. While the company's annual sales have been steadily increasing, this growth has been the result of successful “fashion items” that boost sales for only one season. As a result, sales vary from year to year by up to +/- 30%. In addition, most of the revenue comes in during the pre-Christmas season and comes from large toy stores, which pay their orders within 60 days. The expected sales for the coming year are shown below.

To mitigate the risk associated with the seasonal fluctuations of the business, we currently plan production based on the principle of job production, order-on-demand. This allows the company to quickly respond to customer demand but can be more costly than year-round production as a large number of interns / seasonal workers need to be hired and trained. Due to the increasing competitive pressure, we want to discuss the extent to which it makes sense to change our production strategy in order to reduce costs.

When switching to year-round production, we have to consider several factors. The major ones:

• Financing: The company currently has only a limited cash balance and therefore uses a credit line for additional operating capital. This is limited to 1.9 million without further negotiations with the bank and must have a zero balance for at least one month each year. Any required increases in working capital due to the change in production planning must be taken into account.
• Storage costs: Switching to even production will certainly reduce the hiring costs associated with a seasonal workforce, but it needs to be checked how this compares to the costs of holding additional inventories.
• Uncertain demand: Perhaps the biggest challenge is the dependence of sales growth on fashion toys/seasonal items. If we develop a toy that turns out to be a flop, a large amount of worthless inventory is left with even production.

As part of this experiment, we don't have enough information to assess the risk of fashion toys for our business, but there's enough information to investigate the first two points mentioned above.

Play Time's current production approach is to wait for a customer order and then push that order through the production system as quickly as possible. This approach requires the company to have considerable excess capacity that is only used during the Christmas season, and temporary workers need to be hired and trained each year who can meet quality standards. To break out of this loop, Play Time is considering moving to a steady production where the company produces at a relatively steady pace throughout the year and builds up inventory that can be sold during the peak season. Figure 2 compares the impacts of these two approaches on Play Time's cash, inventory and credit line/liabilities with the bank.

The fundamental differences between the two approaches are easily seen in Figure 2. In the first scenario, where we only produce when we have an order, inventories remain constant. In the balanced production environment, inventories accumulate until they are sold during the Christmas season.

Also, the impact on the required operating capital is immediately apparent. In the case of build-to-order, our credit line is only needed during the high season, when customers place large orders but pay only 1-2 months later.

In the scenario with balanced production, the credit line is already drawn on in the third month, the limit is exceeded in the sixth month, and it is unclear whether we will be able to meet the bank's requirement to hold a zero balance for at least one month per year. So if we want to switch to a flat, balanced production, so much capital is tied up in the inventories that we would have to renegotiate the conditions for the credit line with the bank. Does it make sense for us to apply for a higher credit limit? And if so, how do we get our bank advisor to sign this application?

At first glance, our situation seems hopeless. In reality, our perspective is just a little bit limited because we are considering two extreme scenarios that are opposed to each other.

Currently, everything is produced to order, no stocks are held. However, we would like to have an environment where production takes place evenly throughout the year while reserves are held simultaneously. But what about all the scenarios that lie between these two extremes? What if production ramps up three months before the peak season? How much of the cost savings from the higher utilization and the lower dependence on the seasonality of production would this bring? The Monte Carlo method can be applied to this production scenario of Play Time to evaluate the effects of different production schedules on the profitability of the firm.

How can one look into the future?

Based on Play Time's balance sheet and revenue forecast, the company's finances, such as the need for working capital and inventory, can be calculated.

We assign a random weight of the production quantity to each month. The results are summarized in Figure 3. The two extremes mentioned above are marked. Each random result is shown on the coordinate system. Here, net profit on the left, utilization of our credit line below and our warehouse throughput on the right, are considered factors.

The expected financial performance and need of Play Time Toy Co for a range of possibilities.

Based on these results, we can see that there are a multitude of production plans between the two options we are considering. However, also that our net profit is tied to how high our throughput is in our warehouse.

It's interesting to observe the boundary at which net profit for a given level of credit usage is maximized (the so-called efficient frontier is also an exciting concept when it comes to how you set up your stock portfolio). So we have to find out that our idea of ​​balanced production is far from this border. Our beautiful plan to produce evenly all year long is therefore neither optimal nor efficient for the operating result, even if we are able to negotiate again with the bank. What other options do we have?

Apart from pure build-to-order, there are many options that reduce the seasonality of production but do not require renegotiations with the bank. In the figure below you can see the different monthly scenarios in a comparison of Cost of goods sold (COGS) and production level.

We find: make to order outperforms all other options in terms of profitability and unit economics. It is also clearly evident that the need for working capital increases the earlier production is shifted in the year. Since this capital comes from the credit line, the interest payments narrow the profit margin and reduce the overall profitability of Play Time.

In production planning, we have to evaluate a series of possible scenarios to find the approach that best aligns the goals of our company. In particular, there is a limit at which the return on invested capital is maximized, so we should make sure that we consider this limit when we think about how much we want to pre-produce or store. In the case of Play Time, the need for working capital would exceed the credit line provided by the bank with further growth.

What ERP systems can I use Predictive Analytics with?

So we see that we can use our Predictive-Analytics system for forecasting to plan certain scenarios and thus be able to react a little more flexibly to unforeseeable situations. Microsoft Dynamics 365 Sales offers in the packages Dynamics 365 Sales Premium and Dynamics 365 Sales Enterprise as a Predictive-Analytics-ERP system for example the feature “sales forecasts” which allows you to do the following things:

• Identify sales pipeline risks
• Track the individual sales performance of your sales team based on quotas, and thus be able to offer proactive coaching.
• Use the trends and forecasts to anticipate your sales or revenue results and if necessary reallocate resources.
• Use projected estimates to adjust the product strategy or give your investors updated projections

But also for your marketing, Microsoft Dynamics 365 Business Central as a Predictive-Analytics-ERP system can help you, for example, to distribute lead scores and thereby also increase the efficiency of your media mix, or your marketing activities.

In Dynamics you can set under Settings > Lead Management > Scoring Configuration. The Automated Marketing's Qualification, that allows you to partially automate the lead-to-opportunity process based on a score generated by the lead scoring model.

In short Microsoft Dynamics 365 with Predictive Analytics can therefore: help you say goodbye to your forecasting spreadsheet, make your marketing and sales process more efficient and automated with the help of the learning forecasting function, help you distribute your resources better, adapt your strategy to unforeseen circumstances, offer important support and insights to the salespeople who need it; your investors, your CEO, or your Head of Marketing will certainly be thrilled when you give them a “look into the crystal ball” and deliver automated, robust forecasts. Templates for this can be found here.

In this blog post from Microsoft, you can learn how to use Power BI and Dynamics 365 to fully exploit the potential of Predictive-Analytics-ERP systems. Microsoft Dynamics 365 Business Central is a standard software for ERP systems and is available as a business management solution in the cloud or on-premises. Companies can network their operational processes across departments with the expandable software. The functions are categorized by financial management, sales and service management, project management, supply chain management, production and merchandise management, reporting and real-time analytics.

Predictive Analytics ERP made in Germany

However, there is also Predictive Analytics made-in-Germany. SAP HANA Cloud and SAP Analytics Cloud can support you in creating a forecast for your market share. Using machine learning/predictive analytics, the accuracy of market share forecasts is improved and enables you to segment very granularly. In a dashboard, users can create accurate forecasts and quickly see trends and revenue flows.

The benefits for your team?

• Less manual effort and Excel tables
• More precision in your forecasts since these are no longer susceptible to “human error”
• Easy access to the information, which are clearly presented in a dashboard
• Recyclable project setup that you can also reuse for other use cases

In a 10 part blog post series you can see very granularly how you can integrate Predictive Analytics ERP into your running SAP setup. SAP S/4HANA describes itself as an 'intelligent ERP system' with integrated AI and machine learning. The software is supposed to be able to transform and make business processes more efficient through intelligent automation and runs in SAP HANA (in-memory database). The functions are categorized into asset management, finance, manufacturing, R&D and design, sales, service, purchasing and supply chain. The software is available on-premise, in a public or private cloud, or in a hybrid environment. Here are all the reviews and testimonials.

What other ERP systems offer Predictive Analytics?

Bilbee as one of the winners of the awards “Top Rated ERP Systems” and “Leader ERP Systems” Q1/22 definitely belongs on this list. The functions of Billbee include order processing with (automatic) creation of order documents, payment reconciliation, shipping processing or even sending emails to customers. Furthermore, users can centrally manage and align their article and inventory levels across channels. A flexible automation supports the control of workflows and processes, whereby the automation rules can be individually adapted to your own needs. According to their own statements, over Billbee 10,000 dealers use monthly. The monthly price of the self-service tool is based on the number of orders. Here's the review..

Oracle NetSuite is a cloud-based ERP software that should help its users achieve growth and drive innovation. The main functions of the software include financial management, order management, manufacturing management, supply chain management, warehousing, inventory management and procurement. In addition, NetSuite includes integrated business intelligence in which data is collected in a dashboard and actionable business insights can be extracted. Oracle NetSuite can be supplemented and customized with functions as the company grows. For example, through the Integrated Budgeting, Planning and Forecasting Module which allows you to make reports and analyses like in the above example. Here goes to review.